Question 180 A quarterly report has tables whose cell meanings depend on row headings and footnotes; embedding each numeric cell independently loses context. What prevention measure is most targeted to this underlying cause? Make source/version keys mandatory and enforce deterministic chunk IDs in the Delta ingestion schema. Require inherited heading metadata and region-specific retrieval regression queries. Use table-aware extraction and unit/footnote association tests in the ingestion pipeline. Preserve step-tree ancestry and branch-qualified query tests during structured document ingestion. ExplanationCorrect Answer C:Table values without labels are semantically ambiguous and invite unsupported numeric explanations. Other Answers A:This control addresses a different exposure: An ingestion notebook embeds policy chunks directly but does not store their source document IDs, making outdated vectors impossible to revoke. It would not prevent this failure, because table values without labels are semantically ambiguous and invite unsupported numeric explanations. B:That table values without labels are semantically ambiguous and invite unsupported numeric explanations. D:Table values without labels are semantically ambiguous and invite unsupported numeric explanations.
Question 181 A policy manual repeats the same section title across regions, and flat chunks discard chapter ancestry so search confuses two regions. Which implementation decision best restores the intended behavior? Include the complete hierarchy of region, chapter and section headings with each chunk's text and metadata. Serialize each logical table range with column names, row labels, units and associated footnotes before indexing. Chunk each decision branch with its inherited parent-step condition and stable procedure path rather than flattening all steps into independent text. Apply a relevance re-ranker to the initial candidate set using the full query and candidate passages before choosing generation context. ExplanationCorrect Answer A:The action targets the stated cause: Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. Check the repair by: query an exception tied to one region and verify the indexed hierarchy resolves to that region's section. Other Answers B:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. C:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. D:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures.
Question 182 A policy manual repeats the same section title across regions, and flat chunks discard chapter ancestry so search confuses two regions. Which underlying technical principle explains why this occurs? Table values without labels are semantically ambiguous and invite unsupported numeric explanations. Conditional procedure meaning depends on ancestor branches; isolating a step without its parent condition can produce the wrong diagnostic action. First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. ExplanationCorrect Answer D:Include the complete hierarchy of region, chapter and section headings with each chunk's text and metadata. Other Answers A:That mechanism describes another problem: A quarterly report has tables whose cell meanings depend on row headings and footnotes; embedding each numeric cell independently loses context. It does not explain this case, where headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. B:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. C:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures.
Question 183 A policy manual repeats the same section title across regions, and flat chunks discard chapter ancestry so search confuses two regions. After correction, what targeted test should demonstrate success? Retrieve the amber-LED branch and verify its parent test, branch label and action are present together with source IDs. Query an exceptional metric and verify the returned table chunk includes its unit, row definition and footnote qualifier. Query an exception tied to one region and verify the indexed hierarchy resolves to that region's section. Compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool. ExplanationCorrect Answer C:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. Other Answers A:Query an exception tied to one region and verify the indexed hierarchy resolves to that region's section. B:Query an exception tied to one region and verify the indexed hierarchy resolves to that region's section. D:Query an exception tied to one region and verify the indexed hierarchy resolves to that region's section.
Question 184 A policy manual repeats the same section title across regions, and flat chunks discard chapter ancestry so search confuses two regions. Which release safeguard most directly addresses this failure class? Benchmark two-stage retrieval with separate candidate-recall and final-ranking acceptance thresholds. Require inherited heading metadata and region-specific retrieval regression queries. Use table-aware extraction and unit/footnote association tests in the ingestion pipeline. Preserve step-tree ancestry and branch-qualified query tests during structured document ingestion. ExplanationCorrect Answer B:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. Other Answers A:That headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. C:Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. D:This control addresses a different exposure: A diagnostic flowchart has nested numbered steps with conditional branches, but flat chunks separate an amber-LED branch from its parent test condition. It would not prevent this failure, because headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures.
Question 185 A hybrid search returns the correct maintenance manual at rank twelve, below generic articles that match keywords but not the failure code. Which design change directly satisfies the stated constraints? Use a bounded candidate-recall stage and re-rank only a measured subset that preserves required recall at acceptable latency. Include the complete hierarchy of region, chapter and section headings with each chunk's text and metadata. Filter candidates by user permissions before re-ranking, and independently enforce authorization at the underlying source service. Apply a relevance re-ranker to the initial candidate set using the full query and candidate passages before choosing generation context. ExplanationCorrect Answer D:The action targets the stated cause: First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. Check the repair by: compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool. Other Answers A:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. B:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. C:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals.
Question 186 A hybrid search returns the correct maintenance manual at rank twelve, below generic articles that match keywords but not the failure code. Which causal explanation best accounts for the failure? Headings create disambiguating context; removing ancestors yields near-identical embeddings for distinct jurisdiction-specific procedures. Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. Cross-encoder scoring cost typically grows with query-candidate pairs; indiscriminate larger candidate lists increase processing time. First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. ExplanationCorrect Answer D:Apply a relevance re-ranker to the initial candidate set using the full query and candidate passages before choosing generation context. Other Answers A:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. B:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. C:That mechanism describes another problem: A team increases the vector search candidate count from 40 to 1,000 while keeping a cross-encoder re-ranker, causing unacceptable response latency. It does not explain this case, where first-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals.
Question 187 A hybrid search returns the correct maintenance manual at rank twelve, below generic articles that match keywords but not the failure code. Which concrete acceptance test best demonstrates the proposed change works? Compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool. Probe with a low-privilege identity and show that restricted IDs never reach the re-ranker input or answer context. Query an exception tied to one region and verify the indexed hierarchy resolves to that region's section. Plot recall and end-to-end latency as candidate count varies, then choose a validated budget under the service SLO. ExplanationCorrect Answer A:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. Other Answers B:Compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool. C:Compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool. D:Compare relevant-document rank and grounded answer citations before and after the re-ranker on the same candidate pool.
Question 188 A hybrid search returns the correct maintenance manual at rank twelve, below generic articles that match keywords but not the failure code. Which operational practice should be made a durable requirement? Require inherited heading metadata and region-specific retrieval regression queries. Benchmark two-stage retrieval with separate candidate-recall and final-ranking acceptance thresholds. Add latency and candidate-count guardrails to the retrieval pipeline with routine regression benchmarks. Treat access filtering as a pre-ranking invariant and test it with negative authorization fixtures. ExplanationCorrect Answer B:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. Other Answers A:First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. C:This control addresses a different exposure: A team increases the vector search candidate count from 40 to 1,000 while keeping a cross-encoder re-ranker, causing unacceptable response latency. It would not prevent this failure, because first-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. D:That first-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals.
Question 189 A re-ranker receives confidential chunks before caller entitlements are checked and places one unauthorized report first in the candidate list. What should the implementation team change first to resolve the described defect? Select sources using publication status and caller entitlements before embedding or retrieval exposure, with private documents in restricted collections. Filter candidates by user permissions before re-ranking, and independently enforce authorization at the underlying source service. Use a bounded candidate-recall stage and re-rank only a measured subset that preserves required recall at acceptable latency. Apply a relevance re-ranker to the initial candidate set using the full query and candidate passages before choosing generation context. ExplanationCorrect Answer B:The action targets the stated cause: Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. Check the repair by: probe with a low-privilege identity and show that restricted IDs never reach the re-ranker input or answer context. Other Answers A:Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. C:Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. D:Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering.
Question 190 A re-ranker receives confidential chunks before caller entitlements are checked and places one unauthorized report first in the candidate list. Why does the current approach fail under these conditions? Cross-encoder scoring cost typically grows with query-candidate pairs; indiscriminate larger candidate lists increase processing time. Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. First-stage recall can include the right answer without placing it in the limited top-k; re-ranking changes ordering using richer query-passage signals. Corpus inclusion and access control determine what evidence a model can see; an answer prompt cannot reliably conceal already exposed confidential text. ExplanationCorrect Answer B:Filter candidates by user permissions before re-ranking, and independently enforce authorization at the underlying source service. Other Answers A:Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. C:That mechanism describes another problem: A hybrid search returns the correct maintenance manual at rank twelve, below generic articles that match keywords but not the failure code. It does not explain this case, where ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering. D:Ranking is not a security control; scoring restricted passages still exposes them to a model or service before permission filtering.
